When you build voice assistants for Amazon Alexa and Google Assistant, most of the control over what the AI is doing and how its performing is lost. You can't access audio files, so you can't really understand what users actually said. You can't see anything about how the NLU classified an utterance, either. With Alexa, you don't even get the transcript of what the user said. All of this lack of visibility results in a lack of control over your voice assistant's performance. That means you can't improve your user experience in line with your actual user behaviour and results in sub-par experiences.

To show you how you can regain control of your voice assistant performance and create delightful, ever-improving customer experiences, we're joined by the team at Zammo.ai, Guy Tonye and Stacey Kyler. Guy and Stacey walk us through some of the strategic changes that have occurred when it comes to deploying voice assistants in the enterprise. Changes brought about by, you guessed it, COVID-19.

They speak to why early deployment and quick learning is better than big up-front, waterfall-style deployments, and share their approaches to understanding your assistants performance, and methods of improvement.

We dive into why having control over you ASR and NLU is so important, what levers of control are available, how to understand the data you're receiving from your NLU and how you can use things like call audio, transcripts and confidence scores to refine and improve your assistant over time.

About Zammo.ai

The goal of Zammo.ai is to allow everyone to get things done faster through voice assistants and have more free time doing what they love. To make this reality, they built one of the world's leading cross-platform voice app solutions to move the world from screens to voice interactions. It's solution enables organizations to create Voice Apps in minutes, not months.

Links

Zammo.ai

Stacey Kyler on Linkedin

Guy Tonye on Linkedin

Hosted on Acast. See acast.com/privacy for more information.